A COMPARATIVE-STUDY OF 2 SEARCH STRATEGIES FOR CONNECTED WORD RECOGNITION - DYNAMIC-PROGRAMMING AND HEURISTIC-SEARCH

被引：14

作者：

NEY, H

机构：

[1] Philips Research Laboratory Aachen, Aachen

来源：

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE | 1992年 / 14卷 / 05期

关键词：

BEST-1ST SEARCH; BRANCH-AND-BOUND SEARCH; BREADTH-1ST-SEARCH; CONNECTED WORD RECOGNITION; DYNAMIC PROGRAMMING; HEURISTIC SEARCH; MINIMUM COST PATH FOR FINITE STATE NETWORKS; SEARCH FOR CONTINUOUS SPEECH RECOGNITION;

D O I：

10.1109/34.134063

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

A most successful approach to recognizing continuous speech is to model the recognition problem as one of finding an optimal path through a finite state network. This paper presents a comparison of two search strategies for finding the optimal path: dynamic programming and heuristic search. The comparison is based on theoretical considerations and experimental tests on a digit string task.

引用

页码：586 / 595

页数：10

共 13 条

[1]

Bridle J. S., 1982, Proceedings of ICASSP 82. IEEE International Conference on Acoustics, Speech and Signal Processing, P899

[2] DESIGN OF A LINGUISTIC STATISTICAL DECODER FOR RECOGNITION OF CONTINUOUS SPEECH [J].

JELINEK, F ;

BAHL, LR ;

MERCER, RL .

IEEE TRANSACTIONS ON INFORMATION THEORY, 1975, 21 (03) :250-256

[3]

KAMP Y, 1985, SPEECH SPEAKER RECOG, P115

[4] STRUCTURAL METHODS IN AUTOMATIC SPEECH RECOGNITION [J].

LEVINSON, SE .

PROCEEDINGS OF THE IEEE, 1985, 73 (11) :1625-1650

[5]

Lowerre B., 1980, TRENDS SPEECH RECOGN, P340

[6] THE USE OF A ONE-STAGE DYNAMIC-PROGRAMMING ALGORITHM FOR CONNECTED WORD RECOGNITION [J].

NEY, H .

IEEE TRANSACTIONS ON ACOUSTICS SPEECH AND SIGNAL PROCESSING, 1984, 32 (02) :263-271

[7]

NEY H, 1987, APR P IEEE INT C AC, P833

[8]

NEY H, 1982, 6TH P INT C PATT REC, P1119

[9]

Nilsson N.J., 1980, PRINCIPLES ARTIFICIA

[10]

Sedgewick R., 1983, ALGORITHMS

← 1 2 →